This art icle was downloaded by: [ Neeraj Agarwal] On: 12 March 2012, At : 22: 55 Publisher: Taylor & Francis I nform a Lt d Regist ered in England and Wales Regist ered Num ber: 1072954 Regist ered office: Mort im er House, 37- 41 Mort im er St reet , London W1T 3JH, UK Marine Geodesy Publicat ion det ails, including inst ruct ions f or aut hors and subscript ion inf ormat ion: ht t p: / / www. t andf online. com/ loi/ umgd20 Simulated Heat Content Variability in the Upper Layers of the Tropical Indian Ocean Imran M. Momin Suj it Basu a a , Neeraj Agarwal , Abhij it Sarkar a b , Rashmi Sharma & Vij ay K. Agarwal a , Neet u a , c a At mospheric and Oceanic Sciences Group, Space Applicat ions Cent re, Ahmedabad, India b Max Planck Inst it ut e f or Met eorology, Hamburg, Germany c Bhaskaracharya Inst it ut e of Space Applicat ions and Geoinf ormat ics, Gandhinagar, India Available online: 09 Mar 2012 To cite this article: Imran M. Momin, Neeraj Agarwal, Rashmi Sharma, Neet u, Suj it Basu, Abhij it Sarkar & Vij ay K. Agarwal (2012): Simulat ed Heat Cont ent Variabilit y in t he Upper Layers of t he Tropical Indian Ocean, Marine Geodesy, 35: 1, 66-81 To link to this article: ht t p: / / dx. doi. org/ 10. 1080/ 01490419. 2011. 572759 PLEASE SCROLL DOWN FOR ARTI CLE Full t erm s and condit ions of use: ht t p: / / www.t andfonline.com / page/ t erm s- and- condit ions This art icle m ay be used for research, t eaching, and privat e st udy purposes. Any subst ant ial or syst em at ic reproduct ion, redist ribut ion, reselling, loan, sub- licensing, syst em at ic supply, or dist ribut ion in any form t o anyone is expressly forbidden. The publisher does not give any warrant y express or im plied or m ake any represent at ion t hat t he cont ent s will be com plet e or accurat e or up t o dat e. The accuracy of any inst ruct ions, form ulae, and drug doses should be independent ly verified wit h prim ary sources. The publisher shall not be liable for any loss, act ions, claim s, proceedings, dem and, or cost s or dam ages what soever or howsoever caused arising direct ly or indirect ly in connect ion wit h or arising out of t he use of t his m at erial. Marine Geodesy, 35:66–81, 2012 Copyright © Taylor & Francis Group, LLC ISSN: 0149-0419 print / 1521-060X online DOI: 10.1080/01490419.2011.572759 Simulated Heat Content Variability in the Upper Layers of the Tropical Indian Ocean IMRAN M. MOMIN,1 NEERAJ AGARWAL,2 RASHMI SHARMA,1 NEETU,1 SUJIT BASU,1 ABHIJIT SARKAR,1 AND VIJAY K. AGARWAL3 Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 1 Atmospheric and Oceanic Sciences Group, Space Applications Centre, Ahmedabad, India 2 Max Planck Institute for Meteorology, Hamburg, Germany 3 Bhaskaracharya Institute of Space Applications and Geoinformatics, Gandhinagar, India Ocean General Circulation Model (OGCM) simulations from 1970–2007 are used to study the upper ocean heat content variability in the Tropical Indian Ocean (TIO). Model computed heat contents up to 50 m (denoted by HC50 m hereafter) representing upper ocean heat content and 300 m (HC300 m) representing heat content up to thermocline depth are first compared with heat contents computed from observations of two buoys in the TIO. It is found that there is good agreement between the model and observations. Fourier analysis of heat content is carried out in different regions of TIO. The amplitudes of semi-annual variability for HC50 m and HC300 m are observed to be greater than those for the annual variability in the Bay of Bengal, while in the Arabian Sea there is a mixed result. Heat content tendency is known to be governed by net surface heat flux and horizontal as well as vertical heat transports. For understanding the relative importance of these processes, a detailed analysis of these terms in the tendency equation is carried out. Rossby wave is observed in the annual mode of heat transport while equatorial jet and Kelvin waves are observed in the semi-annual mode of heart transport. Finally, the correlation between heat content and sea surface temperature (SST) and sea level anomaly (SLA), taken one at a time, is computed. It is found that the correlation improves significantly when both these quantities are together taken into account. Keywords Heat content, Ocean General Circulation Model, Tropical Indian Ocean, Fourier analysis 1. Introduction It is now widely recognized that ocean plays a major role in driving climate and weather variations primarily through its large heat content. This quantity and the sea surface temperature (SST) are the two most important parameters in air-sea interaction studies. The net heat flux entering into the ocean is either stored locally primarily in the upper layers or is transported horizontally as well as vertically. Accordingly, the heat content of the ocean varies either due to changing air-sea heat flux or to heat transport. Descriptions of the annual Received 9 August 2010; accepted 2 November 2010. Address correspondence to Sujit Basu, Atmospheric and Oceanic Sciences Group, Space Applications Centre, Ahmedabad 380 015, India. E-mail: skbasu@sac.isro.gov.in 66 Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 Heat Content in the Tropical Indian Ocean 67 cycle of the heat content of the global oceans (Levitus 1984; Tourre and White 1995) and of the individual basins such as the Pacific Ocean (Lie and Endoh 1991; Yan et al. 1995) and the Atlantic Ocean (Reverdin et al. 1991; Weingartner and Weisberg 1991) are available in the literature. Geographic distribution of the annual signal and seasonal variation of the heat content in the upper 100 m and 300 m in the western Mediterranean Sea and the eastern Mediterranean Sea have been studied by Picco (1990) and Ibrahim (1993). They compared the zonal annual trend of the monthly mean storage in the Levantine, Aegean and Ionian seas and showed that the heat storage in both layers (100 m and 300 m) is higher in the Levantine Sea than in the Aegean and Ionian seas. Rao and Sivakumar (1998) used global ocean temperature climatology data to study seasonal variation of heat content in the upper layer of the Tropical Indian Ocean (TIO). They showed that the correlation between heat content and SST degrades with depth. They also showed that the annual mode of heat content is weak in the equatorial band while the most prominent annual mode in the uppermost 50 m is in the southern TIO. Moreover, in the uppermost 300 m, the prominent modes are off Arabia and southwest and southeast India. Gnanaseelan et al. (2003) studied the interannual variability of mixed layer heat content in the equatorial India Ocean (EIO) using a simple mixed layer model. In Geophysical Fluid Dynamics Laboratory (GFDL) coupled model simulation, Indian Ocean dipole/zonal mode events have been found to grow through feedbacks between heat content anomalies and SST-related atmospheric anomalies, especially in the eastern EIO (Song et al. 2007). The interannual variability of the heat content of the upper layer in the EIO was evaluated by Polonskii et al. (2007) using XBT data. Empirical orthogonal function (EOF) and rotated EOF analysis of SST and heat content up to 400 m depth (HC400) in the global ocean during 1979–1991 revealed that the dominant EOF modes for the SST display peak values in the summer–fall of 1983 and the summer of 1987 in the Indian Ocean, while peak values of HC400 occur during the fall of 1982 and the summer of 1987 (Tourre and White 1995). The present work revisits some aspects of the heat budget studies carried out in the TIO by earlier researchers and also throws light on some new aspects by making use of the Ocean General Circulation Model (OGCM) simulations of long integration (38 years). In the present study, we seek to diagnose the contribution to heat content variability from the different terms of the heat content tendency equation like the horizontal/vertical heat transport and surface heat flux. A distinctive aspect of the study is the diagnostics of heat content variability at different temporal scales in the entire TIO. The next section describes the model and data used in the study. Results of validating the model against observations are described in section 3. Various processes responsible for the heat content variations are analyzed in section 4, and the major findings are summarized in section 5. 2. Model and Data Used in the Study Modular Ocean Model version 3.1 (MOM 3) (Pacanowski and Griffies 2000) was used in this study. The bottom topography is based on 1/12◦ by 1/12◦ resolution data from U.S. National Geographic Data Center. The horizontal resolution varies from 0.5◦ in the Indian Ocean to 2◦ elsewhere. The model has 35 levels. Out of these, 24 levels are used to define the upper 300 m depth. The first model level is at 2.5 m depth. Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 68 I. M. Momin et al. The model was spun up from rest by using monthly climatological wind stress (Hellerman and Rosenstein 1983) and restoring boundary condition for SST and sea surface salinity. The initial temperature and salinity were taken from Levitus climatology. The model was run for 50 years to achieve steady circulation. Interannual runs were made from 1950 onwards on a daily basis using NCEP reanalysis data. Model outputs from 1970–2007 (38 years) have been used in the present study. The model was forced at the surface with daily mean air temperature at 2 m height, specific humidity of the air at 2 m, net shortwave radiation, net long wave radiation and zonal and meridional wind at 10 m height. NCEP/NCAR reanalysis (Kalnay et al. 1996) datasets are available in 1.875o × 1.875o resolution. The turbulent fluxes (sensible heat flux, latent heat flux) are calculated using bulk aerodynamic formulae. A wind-dependent drag coefficient (Kara et al. 2000) was used. Monthly climatological river (nearly 3,000 rivers) discharge data over global ocean available from UNESCO site were used. TRITON (Triangle Trans-Ocean buoy Network) is a series of buoys for measuring surface atmospheric as well as upper oceanic parameters. The buoys are deployed by Japan Agency for Marine-Earth Science and Technology (JAMSTEC) in collaboration with many countries in and around the Pacific Ocean as part of International Climate Research Program. These buoys measure wind, air temperature, humidity, precipitation, short wave radiation, water temperature, salinity, and current. The water temperature was measured down to 750 m at depths of 1.5, 25, 50, 75, 100, 125, 150, 200, 250, 300, 500,and 750 m. The data from one of these buoys at 1.5◦ S, 90◦ E in the eastern EIO were used in the present study. The errors from uncertainties of wind speed, air temperature and relative humidity are 0.01 m/s, 0.1◦ C, and 2%, respectively, for the TRITON buoy. Surface meteorological and air-sea fluxes and subsurface oceanographic observations for one year (October 1994 to October 1995) from a surface mooring deployed by Woods Hole Oceanographic Institution (Weller et al. 1998) at 15.5◦ N, 61.5◦ E in the Arabian Sea (AS) were also used in this study. The buoy measured sub-surface temperature at very fine vertical resolution. Details of the observations from the two buoys used in the present study, their locations and period are described in Table 1. Table 1 The buoys used in the present study, their locations and period Data source Locations WHOI 16.5◦ N, 61.5◦ E TRITON 1.5◦ S, 90◦ E Vertical resolution of temperature profile (m) 0.17, 0.43, 0.92, 1.37, 1.41, 1.8, 1.91, 2.4, 3.5, 4.5, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 72, 80, 90, 100, 125, 150, 175, 200, 225, 250, 300 m. 1.5, 25, 50, 75, 100, 125, 150, 200, 250, 300, 500, 750 m. Period Oct 1994–Oct 1995 Jan 2001–Dec 2007 Heat Content in the Tropical Indian Ocean 69 The heat content (HC) with respect to a fixed depth z is calculated following Rao and Sivakumar (1998) as:
z Tdz HC = ρCp 0 where the symbols are as follows: ρ—Mean density of seawater (1024 kg/m3) Cp—Specific heat of sea water (4000 J/Kg K) T—Temperature of the seawater (◦ C) as a function of depth Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 Heat content was computed using model temperatures up to the depths of 50 m (HC50 m) and 300 m (HC300 m). Similar computation was done using buoy observations for the purpose of comparison. 3. Results and Discussion 3.1. Comparison of Model HC Against Observations For the purpose of comparison, a particular buoy location and the nearest model grid point have been taken as a collocated pair. Model computed HC50 m and HC300 m have been compared against data from WHOI mooring and TRITON buoy (Figure 1). Figure 1. Comparison of HC50 m and HC300 m calculated from WHOI (61.5◦ E, 15.50 N) and TRITON (90◦ E, 1.5◦ S) buoys and model. 70 I. M. Momin et al. Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 Table 2 Statistics of the comparison of model heat content with buoy heat content Buoys Parameters (J/m2) Correlation RMSE (109 J/m2) Buoy standard deviation (109 J/m2) WHOI[61.5◦ E, 15.5◦ N] WHOI[61.5◦ E, 15.5◦ N] TRITON[90◦ E, 1.5◦ S] TRITON[90◦ E, 1.5◦ S] HC50 m HC300 m HC50 m HC300 m 0.9 0.73 0.74 0.78 0.11 0.96 0.08 0.8 0.25 1.2 0.09 1.0 It can be seen that the model simulated HC50 m and HC300 m are in good agreement with buoy data at both the locations. The only exceptions to this are for model HC300 m during October–December 1994 and July–August 1995 and for model HC50 m in October–November 1994 and July–August 1995, both in the central AS. Model HC300 m do not agree with the observation in the central AS due to strong vertical displacement of the thermocline (Weller et al. 2002) not reproduced by the model properly. The vertical mixing in the model is not able to capture such high vertical displacement of the thermocline. Regarding HC50 m, the discrepancy in these months could be attributed to the inaccuracy of the NCEP reanalysis winds in this region. This inaccuracy in forcing gave rise to inaccuracy in model simulated heat content. More recently, the accuracy has increased after assimilating QuikSCAT winds. Hence the simulations have turned more accurate and consequently are in better agreement with observations. In Table 2 we present the statistics of the comparison of model computed HC50 m and HC300 m with similar quantities computed using data from the buoys. Correlations are greater than 0.70. It can also be seen that the root-mean-square errors (RMSE) are less than the standard deviations of buoy data. 3.2. Seasonal March of Heat Content in the TIO HC50 m values, averaged bi-monthly over a 38-year period (1970–2007), are shown in Figure 2. High heat content (HC50 m) in the northern Indian Ocean in May–June is attributed to the influence of intense heating. The net heat flux (Figure 4) is positive and increases from March to May in the northern Indian Ocean. There is a net annual heat gain of the order of 25.3 W/m2 in the AS (Figure 4), which is in agreement with 24 W/m2 suggested by Duing and Leetmaa (1980). Western and northern AS show minima in HC50 m during southwest monsoon season (June–August) and winter (December–February). The winter minima are suggested by the negative value of net heat flux (Figure 4). Minima during June–August are due to the strong monsoonal winds causing coastal upwelling near Somali (Duing and Leetmaa 1980). Weller et al. (2002) suggested that the evolution of 0–40 m heat content is in good agreement with surface heat flux except for June and July in the central AS. The heat content in the upper 50 m in Bay of Bengal (BOB) (Figure 2) is high throughout the year, except for northern region during winter where it is found to be less due to cold/dry winter winds leading to strong evaporation in this region. In southern TIO, contours of HC50 m are aligned zonally throughout the year due to the seasonal march of solar declination. HC300 m averaged bi-monthly over a 38-year period (1970–2007) is shown in Figure 3. In the equatorial region, HC300 m is found to be low throughout the year which is in contrast Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 Heat Content in the Tropical Indian Ocean 71 Figure 2. Bimonthly distribution of HC50 m (108 J/m2) in the Tropical Indian Ocean obtained from model hind cast for 1970–2007. to the HC50 m. The reason for low heat content is the shallow and sharp thermocline. However, in the eastern equatorial region, there is high value of heat content during monsoon transition months (May and November). The equatorial jets and downwelling Kelvin waves cause convergence and associated deepening of thermocline (Wyrtki 1973; Reverdin 1985), resulting in high value of heat content in this part. A band of extremely low HC300 m occurs just south of the equator in the west throughout the year. The reason is Ekman divergence of the clockwise gyre, which is restricted on the south by south equatorial current and on the north by the equatorial countercurrent during winter and by the equatorial jet during monsoon transition (Cutler and Swallow 1984). In the northern AS, HC300 m is found to be high throughout the year and this is due to the deep thermocline. There is significant seasonal variability seen in BOB and more so in the southwestern region. In the southern TIO between 15o to 20◦ S, zonal band of high value of HC300 m is present throughout the year with little seasonal variability. Convergence caused by a westward flowing south equatorial current on the north and eastward flowing South Indian Ocean current on the south (Stramma 1992) is the cause of this high HC300 m. In the AS, the high value of HC300 m observed in the southwest of India and spreading toward the Somali coast during February/March is due to Rossby waves radiating from west coast of India and propagating westward along AS (Jensen 1991; McCreary et al. 1993). In BOB, the model heat content is high during February and March. This is caused by coastal Kelvin waves propagating around the Sri Lankan coast (McCreary et al. 1993). One can clearly observe strong seasonal variability in HC300 m towards western BOB. This is attributed to the seasonally reversing East India coastal currents. Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 72 I. M. Momin et al. Figure 3. Bimonthly distribution of HC300 m in the Tropical Indian Ocean obtained from model hind cast for 1970–2007. 3.3. Rate of Change of Heat Content Variations In Figures 5 and 6 we present the seasonal variation of the rate of change of HC50 m and HC300 m. During February to May, heat content rate is positive in the AS and BOB in the upper ocean (up to 50 m) because of the net heat gain at the ocean surface, as seen from Figure 4. The rate is negative in the AS during southwest monsoon (Figure 6) although net heat flux is positive (Figure 4). Duing and Leetma (1980) found that there was a net heat gain during the southwest monsoon, which was offset by horizontal advection and upwelling resulting in negative rate of change. Satellite imagery shows that cool water upwelled along the Somali and Arabian coasts during the southwest monsoon was advected offshore (Brown et al. 1980; Cagle and Whritner 1981). Buoy measured data in the AS indicated that there is net surface heating during summertime southwest monsoon (Weller et al. 1998). However, surface heat climatologies (Hastenrath and Lamb 1979; Oberhuber 1988) and observations (Rao 1986) suggested that the surface heat flux was negative (a net heat loss in ocean), driving convective entrainment and additional cooling. Weller et al. (2002) show that there is a net heat loss (19.7 W/m2) from the ocean during northeast monsoon in the central AS and that evaporation was the driving component. The buoyancy flux across the sea surface was driven by both heat and mass flux. In the BOB, upper ocean heat content is largely driven by net heat flux at the ocean surface. However, one can see an exception during July–August when, although the net heat at the ocean surface is positive, the upper ocean is losing heat. The reason could be the monsoonal current taking away the heat from the interiors of the BOB towards the south. Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 Heat Content in the Tropical Indian Ocean 73 Figure 4. Bimonthly distribution of net heat flux (W/m2) in the Tropical Indian Ocean. In other months, the variations can be well explained in accordance with the surface heat flux. In the southern TIO, the rate of change of upper ocean HC is primarily governed by the seasonal migration of sun’s radiation. Net surface heat in the EIO region is positive throughout the year and, as expected, the upper oceanic layers gain heat as evident from Figure 5. The seasonal variation of the rate of change of HC300 m is at its maximum in the central AS and coastal BOB from January to March. This high variability in the coastal Bay is suggestive of the coastal Kelvin wave propagation. The coastal Kelvin wave, triggered in BOB, travels all along the coast and after reaching the southeast Arabian Sea radiates mode-2 Rossby waves propagating to the central AS in March (McCreary et al. 1993). The equatorial jet and downwelling Kelvin wave propagate during transition periods (April–May and October–November) in the eastern EIO. These modulate the heat content through their mass convergence and depression of the thermocline depth (Wyrtki 1973; Reverdin 1985). During the summer monsoon, the heat content rate is found to be maximum in the central AS due to deepening of the surface mixed layer (Rao 1986) and the deepening of the thermocline depth due to negative wind stress curl (Hastenrath and Lamb 1979). 3.4. Fourier Analysis of HC In what follows we show, in Figure 7, the result of Fourier analysis of HC50 m and HC300 m in the TIO, divided into three oceanographically distinct regions: the AS (40–70◦ E, 5–25◦ N), BOB (80–100◦ E, 5–25◦ N), and EIO (40–100◦ E, 5◦ S–5◦ N). In the AS, amplitude of semi-annual variability is stronger than that of annual variability for HC50 m while the opposite is true for the HC300 m. For further diagnosis of the Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 74 I. M. Momin et al. Figure 5. Bimonthly distribution of heat content rate (W/m2) in the upper 50 m obtained from model hind cast for 1970–2007. principal temporal scales of variability of the heat content, Fourier analysis of net heat flux was carried out for the AS, BOB and the EIO (figures not shown). It was found that strong semi-annual variability in the HC50 m in the AS agrees with semi-annual variability of net heat flux. Weller et al. (2002) have reported that variation of 0–40 m heat content is in good agreement with that of surface heat flux in the central AS for the whole year, the only exception being the months of June and July. Thus this Fourier analysis seems to confirm the finding of Weller et al. (2002). The amplitudes of semi-annual variability for HC50 m and HC300 m are observed to be greater than those for the annual for the BOB. However, in the EIO the opposite is observed. In the AS the high amplitudes of net heat flux are found on annual and semi-annual scales, while semi-annual scale variation is stronger compared to annual scale. The reason for this may be the following. The net heat flux for the northern AS and northern BOB are high during the monsoon transitions, that is, March–April and September–October (Figure 4). This results in high variability at semi-annual scale. The amplitude of net heat flux in the EIO is found to be low compared to AS and BOB due to the low distribution of net heat flux. The annual wind field and net heat flux exchange are weakest in the EIO (Rao et al. 1991). 4. Processes In this section we study the relative importance of different parameters/processes associated with the various scales of heat content variability. The heat content tendency is governed by the net surface heat flux, horizontal advection, vertical mixing/entrainment and diffusion. Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 Heat Content in the Tropical Indian Ocean 75 Figure 6. Bimonthly distribution of heat content rate (W/m2) in the upper 300 m obtained from model hind cast for 1970–2007. The evolution of the heat content (HC) can be written as: ∂HC/∂t + [u∂HC/∂x + v∂HC/∂y] + w∂HC/∂z = Q + diffusive terms where ∂H C/∂t is the heat content tendency, u∂H C/∂x + v∂H C/∂y is the horizontal heat advection, w∂HC/∂z is the vertical heat transport and Q is the net surface heat flux, which includes the short wave and long wave radiation, latent and sensible heat fluxes. What follows is the study of the relative impact of heat transport and net surface heat flux on the heat content variability in the upper 50 m and 300 m at different temporal scales. In the previous section, it was reported that HC50 m and HC300 m exhibit marked variability at some selected frequencies corresponding to annual (365-day), semi-annual (180-day) and low period intraseasonal (120-day) scales. In order to understand the processes responsible for variations of HC50 m and HC300 m at these frequencies, HC50 m and HC300 m were subjected to band-pass filter with cut-off periods (lower and upper cut-offs) of 110 and 130 for 120-day cycle, 170 and 190 for semi-annual and 260 and 370 for annual cycles. A similar filtered output for ocean advective flux, surface net heat flux and vertical transports were also generated. In the subsequent subsections we discuss the processes/parameters responsible for giving rise to variabilities in HC50 m and HC300 m at the above-mentioned three periodicities. 4.1. Annual Variability of the Heat Content Figure 8 represents the spatial distribution of the annual harmonics of HC50 m and HC300 m. Net heat flux variability and total heat transport variability (advection + vertical) Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 76 I. M. Momin et al. Figure 7. Fast Fourier Transform of HC50 m (upper panel) and HC300 m (lower panel) during 1970–2007. (a) Arabian Sea (40–70◦ E, 5–25◦ N), (b) Bay of Bengal (80–95◦ E, 5–25◦ N) and (c) Equatorial Indian Ocean (40–90◦ E, 5◦ S–5◦ N). computed over 0–300 m are also shown in the same figure. The high variability of the HC50 m was found in the southern TIO, the northern AS and the northern BOB for annual scale while low in the EIO. The distribution of net heat flux also shows minima in the EIO (Figure 8). Rao et al. (1991) suggest that the annual wind field and net heat flux exchange are weakest in the EIO. At 300 m, dynamical processes such as coastal upwelling and the propagating waves produce large variability in the thermocline of the coastal region in the northern TIO. Two major signals are responsible for the variation of HC300 m in the AS. Jensen (1991) used a three and one-half layer reduced gravity model to observe Rossby waves radiating from the coast of the southwest of India and propagating westward toward the central AS. The coastal Kelvin waves travel along the northern boundary of the Bay and reach the east coast of India in March. McCreary et al. (1993) used a two and one-half layer reduced-gravity model to observe Kelvin waves which propagate around Sri Lankan coast and radiate mode-2 Rossby waves from southwest of India into the central AS. The coastal Kelvin waves are also seen in the heat transport of upper 300 m. In the southern TIO (5◦ S–20◦ S), the strong band of variability is found for annual mode of heat transport in upper 300 m due to Rossby wave propagation. The propagation of Rossby waves is also observed in annual mode of HC300 m. The Rossby waves probably driven by the Pacific through flow are seen in model sea level (Woodberry et al. 1989; Perigaud and Delecluse 1992; Basu et al. 2000). These Rossby waves influence the heat content through variation in the thermocline topography. Rao (1986) found that there is negative correlation between Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 Heat Content in the Tropical Indian Ocean 77 Figure 8. Distribution of annual harmonic amplitude for HC50 m (108 J/m2) and HC300 m. Also shown are net heat flux (W/m2) and heat transport (W/m2) in the TIO for annual harmonic amplitude. SST and heat content in upper 200 m layers. This is caused by the Rossby waves propagating in the southern TIO between the Equator and 15◦ S. These waves influence the HC by modulating the thermocline depth. 4.2. Semi-annual Variability of the Heat Content The spatial distribution of the semi-annual harmonic of HC50 m and HC300 m are shown in Figure 9. Net heat flux variability and the total variability of heat transport (advection + vertical) calculated over 0–300 m are also shown in the same figure. The semi annual variability of the HC50 m is most dominant in northern AS and northern BOB. This region also shows maximum variability in net heat flux. The reason is that the northern AS and the northern BOB receive the highest net heat flux during the monsoon transitions, that is, March–April and September–October (Figure 4). These result in highest heat flux in May and November. The variability of HC300 m was strong in the eastern EIO due to mass convergence caused by equatorial jet and downwelling Kelvin waves in both monsoon transitions, during March–May and September–November (Wyrtki 1973; Reverdin 1985). The equatorial jets and downwelling Kelvin waves are observed in semi-annual mode of heat transport. A variability such as letter C is observed in HC300 m and heat transport in upper 300 m. This C-shaped pattern indicates the effects of the western boundary in reflecting Rossby waves (Le Blanc and Boulanger 2001). Lee-Lueng Fu (2006) shows that the semiannual variation of sea surface height (SSH) from the Jason and the TOPEX/Poseidon radar altimeters in the EIO is also characterized by propagating Kelvin waves and Rossby waves. The main features such as equatorial Kelvin waves and off-equatorial Rossby waves are Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 78 I. M. Momin et al. Figure 9. Distribution of semi-annual harmonic amplitude for HC50 m (108 J/m2) and HC300 m. Also shown are net heat flux (W/m2) and heat transport (W/m2) in the TIO for semiannual harmonic amplitude. reported by O’Brien and Hurlburt (1974) in their model simulations. The phase of the mode reveals westward propagation in the regions off the equator and eastward propagation along the equator. These features suggest the roles of Kelvin waves on the equator and Rossby waves off the equator. 5. Correlation Between SST, SLA and Heat Content Following Rao and Sivakumar (1998), correlation between SST and heat content has been studied using the model simulations. Attempts have also been made to study the correlation between heat content and sea level anomaly (SLA) as well as the multiple correlation between heat content and SST and SLA. The results are displayed in Figure 10 in which the upper panel shows the correlations for 50 m and the lower panel shows the correlations for 300 m depth. The correlation between SST and HC50 m is very high in the southern Indian Ocean and reasonably high in the Arabian Sea and BOB, while there are regions of poor correlation in the region around 10◦ S and in the eastern equatorial Indian Ocean. The reason for high correlation in the southern and western TIO is that these are the regions of strong surface wind field with deep and diffuse thermocline. In conformity with the result of Rao and Sivakumar (1998), the correlation deteriorates with depth and there is poor correlation between SST and HC300 m. Interestingly, the result is quite opposite when the correlation between heat content and SLA is studied. There is poor correlation between SLA and HC50 m almost everywhere, whereas this correlation is very strong between SLA and HC300 m, barring isolated pockets. The correlation is poor in the southeastern Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 Heat Content in the Tropical Indian Ocean 79 Figure 10. Correlation of HC50 m (a, b, c) and HC300 m (d, e, f) with SST, SLA, and SST and SLA taken together. From left to right are shown the correlations with SST, SLA, and SST and SLA taken together. TIO for both HC50 m and HC300 m, possibly because southwestward propagating Rossby waves seen in this region (Perigaud and Deelecluse 1992; Basu et al. 2000) influence the heat content through modulation of thermocline topography. Interestingly the correlation between HC300 m and SST is also poor in this region and for the same reason. It was thus felt that the correlation can possibly be improved if one considers both SST and SLA Accordingly, the multiple correlation between the heat content and SST and SLA was studied. As seen from the rightmost panels, there was a significant improvement in the correlation throughout the TIO. Interestingly, both these quantities (SST and SLA) are easily available from satellite measurements. Hence heat content could, in principle, be derived from satellite measurements using simple multiple regression, without going through the elaborate and extensive model calculations. 6. Conclusion Tropical Indian Ocean is an important ocean for its influence on climate and for its importance in air-sea interaction studies. Heat content of the upper ocean is the most important parameter in these studies. Although there have been attempts in the past to compute this quantity using climatological and scattered observations, a systematic study using long period of ocean circulation model simulations is lacking. In the current study efforts have been made to carry out such a study using long periods (38 years) of ocean general circulation model simulations. The simulations of upper ocean heat content have been first validated against observations and the seasonal variability of the heat content in the upper 50 m and 80 I. M. Momin et al. Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 upper 300 m has been described. In order to find out the dominant modes of variability, Fourier analysis has been done. To avoid any data gaps, the continuous time series has been used and the periods corresponding to extreme events such as cyclones have not been filtered out. The major findings are that the dominant modes of variability are annual and semi-annual. Heat content tendency is governed by the net surface heat flux, horizontal advection, vertical mixing/entrainment and diffusion, and the relative importance of these processes has been studied by analyzing these terms. Finally, the correlation between heat content and SST and SLA taken one at a time has been computed. It has been found that the correlation improves significantly when these quantities are taken together to compute the correlation. Since both these quantities are easily measured by ocean observing satellites, it has been conjectured that the heat content can also be found from these quantities via simple multiple correlation. Acknowledgements TAO Project Office of NOAA/PMEL is gratefully acknowledged for RAMA/TRITON data. The authors also thank Geophysical Fluid Dynamics Laboratory (USA) for providing the MOM code and related datasets. NCEP reanalysis data were downloaded from http://cdc.noaa.gov/ References Basu, S., S. D. Meyers, and J. J. O’Brien. 2000. Annual and interannual sea level variations in the Indian Ocean from TOPEX/POSEIDON observations and ocean model simulations. J. Geophys. Res. 105:975–994. Brown, O. B., J. G. Bruce, and R. H. Evans. 1980. Evolution of sea surface temperature in Somali basin during SW monsoon of 1979. Science 209:595–597. Cagle, B. J. and R. Whritner. 1981. Supplement to AS project of 1980; composite of infrared images. Technical Report ONRWEST 81-5SUP. Pasadena, CA: Office of Naval Research, Western Regional Office. Cutler, A. N. and L. C. Swallow. 1984. Surface currents of the Indian Ocean (to 25◦ S, 100◦ E) compiled from historical data archived by the meteorological office. UK, Report No. 187. Bracknell, UK: I.O.S., Wormley. Duing, W. and A. Leetma. 1980. Arabian Sea: Cooling a preliminary heat budget. J. Phys. Oceanogr. 10:307–312. Fu, L.-L. 2006. Intra-seasonal variability of the EIO observed from sea surface height, wind and temperature data. J. Phys. Oceanogr. 37:188–202. Gnanaseelan, C., J. S. Chowdhary, A. K. Mishra, and P. S. Salvekar. 2003. Interannual variability of upper ocean heat content in the northern Indian Ocean. J. Ind. Geophys. Union 7:193–200. Hastenrath, S. and P. Lamb. 1979. Climatic atlas of the Indian Ocean. Part 2: The oceanic heat budget. Madison, WI: University of the Wisconsin Press, 10400. Hellerman, S. and M. Rosentein. 1983. Normal monthly stress over the world ocean with error estimates. J. Oceanogr. 13:1093–1104. Ibrahim, A. M. 1993. Heat storage in the eastern Mediterranean. J. Phys. Oceanogr. 23:1259–1264. Jensen, R. G. 1991. Modeling the seasonal undercurrents in the Somali current system. J. Geophys. Res. 96:22151–22167. Kalnay, E., M. Kanamitsu, R. Kistler, W. Collins, D. Deaven, L. Gandin, M. Iredell, S. Saha, G. White, J. Woollen, Y. Zhu, A. Leetma, R. Reynolds, M. Chelliah, W. Ebisuzaki, W. Higgins, J. Janowiak, K. C. Mo, C. Ropelewski, J. Wang, R. Jenne, and D. Joseph. 1996. The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Met. Soc. 77:437–471. Kara, A., P. A. Rochford, and H. E. Hurlburt. 2000. Efficient and accurate bulk parameterizations for air-sea fluxes for use in general circulation models. J. Atmos. Ocean Technol. 17:1421–1438. Downloaded by [Neeraj Agarwal] at 22:55 12 March 2012 Heat Content in the Tropical Indian Ocean 81 Le Blanc, J. L., and J. P. Boulanger. 2001. Propagation and reflection of long equatorial waves in the Indian Ocean from TOPEX/Poseidon data during the 1993–1998 period. Clim. Dyn. 17:547–557. Levitus, S. 1984. Annual cycle of the temperature and heat storage in the world ocean. J. Phys. Oceanogr. 14:727–746. Lie, H. J. and M. Endoh. 1991. Seasonal to inter-annual variability in the temperature of the upper layer of the northwest Pacific, 1964–1983. J. Phys Oceanogr. 21:385–397. McCreary, J. P., P. J. Kundu, and R. L. Molinari. 1993. A numerical investigation of dynamics, thermodynamics and mixed layer processes in the Indian Ocean. Prog. Oceanogr. 31:181–244. Oberhuber, J. M. 1988. An atlas based on the COADS data sets: The heat budgets of heat, buoyancy and turbulent kinetic energy at the surface of the global ocean. Max-Planck-Institute for Meteorology Report No. 15 Hamburg, Germany: Max-Planck-Institüt fur Meteorologie. O’Brien, J. J. and H. E. Hurlburt. 1974. Equatorial jet in the Indian Ocean: Theory. Science 184:1075–1077. Pacanowski, R. C. and S. M. Griffies. 2000. MOM 3.0 manual. Introduction to MOM and its use. Tech. Rep. 4. Princeton, NJ: GFDL Ocean Group. Perigaud, C. and P. Delecluse. 1992. Annual sea level variations in the southern TIO from GEOSAT and shallow water simulations. J. Geophys. Res. 97:20169–20178. Picco, P. 1990. Climatological atlas of the western Mediterranean. La Spezia, Italy: ENEA, Centro Ricerche Ambiente Marino. Polonskii, A. B., G. Meyers, and A. V. Torbinskii. 2007. Interannual variability of the heat content of the upper layer in the EIO and the Indian-Ocean dipole. J. Phys. Oceanogr. 17:129–140. Rao, R. R., R. L. Molinari, and J. F. Festa. 1991. Surface meteorological and subsurface oceanographic atlas of the tropical India Ocean. NOAA Tech. Memor. ERL-AOML-69. Miami, FL, USA: United States Department of Commerce. Rao, R. R. 1986. Cooling and deepening of the mixed layer in the central Arabian Sea during MONSOON-77: Observations and simulations. Deep Sea Res. 33:1413–1424. Rao, R. R. and R. Sivakumar. 1998. Observed seasonal variation in heat content in the upper layers of the TIO from a new global ocean temperature climatology. Deep Sea Res. 45:67–89. Reverdin, G. 1985. Convergence in the equatorial surface jets of the Indian Ocean. J. Geophys. Res. 90:11741–11750. Reverdin, G., P. Raul, and G. Y. du Penhoat. 1991. Vertical structure of the seasonal cycle in the central equatorial Atlantic Ocean; XBT sections from 1980 to 1988. J. Phys. Oceanogr. 21:227–291. Song, Q., G. A. Vecchi, and A. J. Rosati. 2007. Indian Ocean variability in the GFDL coupled climate model. J. Climate, 20, doi: 10.1175/JCLI4159.1. Stramma, L. 1992. The south Indian Ocean current. J. Phys. Oceanogr. 22:421–430. Tourre, Y. M. and W. B. White. 1995. ENSO signals in global upper ocean temperature, J. Phys. Oceanogr. 25:1317–1332. Weingartner, T. J. and R. W. Weisberg. 1991. A description of the annual cycle in sea surface temperature and upper ocean heat content in the equatorial Atlantic. J. Phys. Oceanogr. 21:83–96. Weller, R. A., A. S. Fischer, D. L. Rudnick, et al. 2002. Moored observations of the upper ocean response to the monsoons in the AS during 1994–1995. Deep Sea Res. Part 2, 49:2195–2230. Weller, R. A., M. F. Baumgartner, S. A. Josey, A. S. Fischer, and J. C. Kindle. 1998. Atmospheric forcing in the AS during 1994–1995; observations and comparisons with climatology and models. Deep Sea Res. Part 2, 45:1961–1999. Woodberry, K. E., M. W. Luther, and J. J. O’Brien. 1989. The wind driven seasonal circulation in the southern TIO. J. Geophys. Res. 94:17985–18002. Wyrtki, K. 1973. An equatorial jet in the Indian Ocean. Science 181:262–264. Yan, X. H., P. P. Niiler, S. K. Nadiga, R. H. Stewart, and D. R. Cayan. 1995. Seasonal heat storage in the north Pacific: 1976–1989. J. Phys. Oceanogr. 100:6899–6926.